Owing to its high time bandwidth product and low probability of intercept,linear frequency modulated signal is widely used in modern radar. Based on wavelet ridge extracting of linear frequency modulated(LFM)signal,a new method used to detect and estimate the LFM signal is presented. It computes the wavelet transform of the signal firstly to obtain the 2D time-scale image,and then extracts the wavelet ridge of the signal with ridge extracting algorithm,finally performs the Hough transform in the time-frequency plane. The detection of lines in the time-frequency plane then reduces to a 2D searching for the peaks in the Hough transform parameter domain. Computer simulation results indicate that the method can detect the LFM waveforms in presence of strong noise effectively.